Knowledge graphs, vast networks of interconnected facts, are crucial for AI applications. But they're often incomplete or inaccurate. Imagine trying to understand the importance of a movie star in a knowledge graph where some movies they acted in are missing details, or even worse, have the wrong director listed. That's where the power of Large Language Models (LLMs) comes in. Researchers have developed a new method called LLMs Empowered Node Importance Estimation (LENIE) that uses LLMs to fill in these gaps and enhance the semantic information within knowledge graphs. LENIE works by first cleverly extracting key facts related to a specific node in the graph. Instead of just randomly picking connections, it uses a clustering technique to select a diverse set of relationships that best represent the node's overall context. Think of it like choosing the most representative scenes from a movie to summarize the plot. This information is then fed to an LLM as a prompt, along with any existing description text. The LLM then generates a richer, more accurate description of the node, effectively supercharging the knowledge graph with more accurate and complete information. Experiments show that this approach significantly improves the ability to estimate the importance of nodes within the graph, especially in scenarios with missing or incomplete data. For example, in a music knowledge graph where only artist names were available, LENIE dramatically improved the accuracy of estimating artist popularity by filling in missing details about their albums, collaborations, and musical styles based on the relationships in the graph. This opens up exciting possibilities for building more robust and accurate knowledge graphs that can power more effective AI applications, from recommendation systems to drug discovery.